1404
NOTES
YOL.
28
The Forma tion of Hexachlorocyclopropane by the Addition of Dichlorocarbene to Tetrachloroethylene
hexachlor~propane.~Subsequently, Davis and Whaleys by chlorinahave prepared 1,1,1,3,3,3-hexachloropane followed by fi,action of 1,1,1,3,3-pentachloropropane tionation to separate the two hexachloropropanes which are formed and have reported that 1,1,1,3,3,3-hexaWILLIAM R. MOORE,S. E. KRIKORIAN, A N D JOSEPH E. LAPRADE chloropropane is a liquid, b.p. 205O, m.p. -27". Since all four hexachloropropanes have Raman spectrae Department of Chemistry, Massachusetts Instztute of Technology, which are consistent with the assigned structures, we Cambrzdge 39, Massachusetts believe that Stevens probably did succeed in preparing hexachlorocyclopropane. Received A'ovenzber SO, 1968 The fact that perchlorocyclopropane was formed a t all in the present study, albeit in very low yields, takes The addition of dihalocarbenes to olefins discovered on greater significance when one considers that dichloroby Doering and Hoffmann' has provided an exceptioncarbene apparently fails to react with ethylene3 (in ally useful and widely used synthesis of dihalocyclopropreference to reaction with t-butoxide). Tetrachloropanes. Studies by Skel12and Doering3have established ethylene is far less reactive than ethylene in typical that dibromocarbene and dichlorocarbene behave as electrophilic reactions (e.g., addition of bromine). electrophilic reagents since electron-donating groups Thus it is possible that in reacting with tetrachloroon an olefin facilitate addition. Accordingly, a highly ethylene, dichlorocarbene may be exhibiting either negatively substituted olefin would be expected to he radical or nucleophilic character. quite unreactive3 toward dihalocarbenes. I n particular, it appeared to be of interest to examine the addition of dichlorocarbene to tetrachloroethylene, an olefin Experimental which is resistant to electrophilic attack, since the Reaction of Chloroform with Potassium t-Butoxide.-Chloropotential adduct, hexachlorocyclopropane, which might form (36 g.) was added dropwise with stirring over a period of 1 prove to be of intrinsic interest, was reported by Stevens4 hr. to a mixture of 50.5 g. of potassium t-butoxide (freed of alcohol by heating a t 140" a t l mm.) and 325 g. of freshly distilled to be unavailable by way of chlorination of cycloprotetrachloroethylene. The reaction mixture was cooled intermitpane. To this end, dichlorocarbene was generated in tently with an ice bath. The mixture was stirred a t room temthe presence of tetrachloroethylene by treatment of perature for an hour and then was poured into water. The chloroform with strong bases' and by the thermal deorganic layer was separated, dried, and the bulk of the tetraconiposition of sodium trichloroacetate5 in l ,2-dichloroethylene was removed by distillation a t atmospheric pressure leaving a dark oil, a portion of which distilled a t 70-120" methoxyethane. Hexachlorocyclopropane was formed (1-2 mm.) leaving a considerable quantity of residual tar. The in these reactions, but in low yields (ca. 0.2-l%).e distillate, shown by gas chromatography to be mainly tetraThe structure of the compound, a white crystalline chloroethylene, was redistilled slowly under reduced pressure, solid, m.p. 104-104.5O, mas established by elemental leaving a solid residue. The latter was sublimed five times a t 40-50" (0.5 mm.) giving 0.21 g. of hexachlorocyclopropane, analysis, which gave an empirical formula of CC12, and m .p . 104.0-104.5 . the determination of the molecular weight cryoscopiAnal. Calcd. for CaC16: C, 14.46; C1, 85.54; mol. wt., 249. cally and by mass spectrometry. The mass spectrum Found: C, 14.46; C1,85.69; mol. wt., 241.10 of C3Cle does not show any peak due to the molecular In solution (carbon disulfide and carbon tetrachloride), perion, but no fragments are formed which have more than chlorocyclopropane shows prominent bands in the infrared a t 850 (s), 905 (m), and 930 (w) cm.-1. Mass spectrum was measured three carbon atoms. The most abundant fragment is on a Consolidated Electrodynamics Corporation Model 21-103C C3C15, corresponding to loss of a single chlorine atom. mass spectrometer with a heated inlet system (140") a t an ionizing Inasmuch as there can be only two compounds with a potential of 70 v. The maEs number of the largest isotopic molecular formula of C3Cls,namely hexachlorocyclopropeak of each monopositive carbon-containing fragment is given as a percentage of the largest peak in the spectrum. Normal isopane and hexachloropropylene, and the latter is a well topic distribution of CIS and C137was observed within each fragknown commercially available liquid, it is clear that the ment (thus all fragments listed contain only CIS except the last solid referred t o above is perchlorocyclopropane. three each of which contains one Cl37). In retrospect, it appears likely that Stevens4did preFragment, mass number (percentage): CC1, 47 (35.4); C&I, pare hexachlorocyclopropane. He reported that 59 (6.4); CaC1, 71 (30.3); CC12, 82 (22.2); C&lz, 94 (12.2); CaC12,106 (15.4); CCl3, 117 (26.4); c&13, 129 (4.4); C3Cla, 141 for seven chlorinationof 1,1,2,2-tetrachlorocyclopropane (16.8); Cd&, 166 (12.8); c3C14, 178 (0.9); CaC15, 213 (100.0). days a t 63' in the presence of ultraviolet light gave Use of Sodium Hydride.-Methyl alcohol (9.6 g., 0.30 mole) hepta- and octachloropropane as the main products, was added dropwise with stirring over a period of 9.5 hr. a t room but also gave a very small amount of a white crystalline temperature to a mixture of 298 g. of tetrachloroethylene, 7.2 g. solid, m.p. 102-102.5O, which on the basis of elemental (0.30 mole) of sodium hydride, and 44.6 g . (0.37 mole) of chloroanalysis he believed to be the then unknown 1,1,1,3,3,3- form. The mixture was processed as above to give 0.42 g. of O
(1) W. von E. Doering and A. K. Hoffmann, J . A m . Chem. Soc., 7 6 , 6162 (1954). (2) P. S. Skell a n d A. Y . Garner. ibid.. 78, 5430 (1956). (3) W. von E. Doering a n d W. A. Henderson, Jr., ibid., 80, 5274 (1958). (4) P. G . Stevens, ibid., 68, (320 (1946). ( 5 ) W. M. Wagner, Proc. Chem. SOC.(London), 229 (1959). (6) (a) Since this work was completed, the formation of hexachlorocyclopropane b y similar means (chloroform and fused potassium hydroxide a t I O R O ) , b u t in higher yields (5-10%). has been reported by S. W. Tobey and It. C. West, Abstracts of Papers presented a t the 142nd National Meeting of the American Chemical Society, Atlantic City, N. J., September 9-14, 1962, p. 95Q; (b) after submission of this manuscript we learned t h a t E. K. Field a n d S. Meyerson have also prepared this compound and measured ita mass spectrum.
hexachlorocyclopropane. When the methyl alcohol was omitted (reflux, 40 hr.) or replaced by t-butyl alcohol (60°, 11 hr,), only a trace of this product was isolated. Use of Sodium Trich1oroacetate.-A mixture of 18.5 g. of sodium trichloroacetate, 80 g. of tetrachloroethylene, and 150 rnl. of 1,2-dimethoxyethane was refluxed 24 hr. After processing in the usual way, the mixture was concentrated by distillation and and both distillate and residue were analyzed by gas chromatog-
(7) Analysis indicated 0.53% hydrogen, a value in between t h e req~rit'rd values of 0.00 for hexachlorocyclopropane and 0.80 for hexach1oror)ropanc. (8) H. W. Davis and A. M. Whaley, J . A m . Chem. Soc.. 1 3 , 2361 (1951). (9) 1%.Gerding and H. G. Haring, Rec. trav. chim., 74, 841 (1955). (10) Determined by Dr. C. M. Starks by freezing point lowering of bennene solutions.
MAY,1963
NOTES
raphy which indicated formation of ca. 1% of hexachlorocylopro~ 6.6).11 The chromatograms showed peaks due to pane ( t = several other minor components, one of which was collected ( t =~ 2.7) and found to have a strong band a t 1763 em.-' but which was not characterized further. In a similar experiment, the gas evolved during the reflux period was collected and analyzed by infrared spectroscopy which indicated that the sample consisted of carbon dioxide that did not contain more than a small amount (< 57") of carbon monoxide. This result apparently precludes significant reaction of dichlorocarbene with trichloroacetate ion to give, over-all, carbon monoxide, trichloroacetyl chloride, and chloride ion by a reaction path formally similar to that describedl2 for the reaction of dichlorocarbene with alkoxide ions. I n other control experiments it was found that the thermal decomposition of sodium trichloroacetate in 1,2-dimethoxyethane did not produce significant quantities of materials with retention times greater than that of the solvent. (11) Retention time relative t o tetrachloroethylene, t~ = 1.00; dimethoxyethane. t~ = 0.33 on silicone oil a t 150'. (12) P. S. Skell and I. Starer, J . A m . Chem. Soc., 81, 4117 (1959).
1,2-
Addition Reactions of rn- and p-Nitronitrosobenzene'
1405
ing alkenes : cyclohexene, cyclopentene, isobutylene, cis- and trans-2-butene. Considering the evidence, this hypothesis concerning the structure of product I1 was then rejected. It was also considered unlikely that the conjugated diene would add two moles of the nitroso compound in a 1,2fashion. It is well established that miziiy aromatic nitroso compounds are in a state of equilibrium between monomer and dimer in s ~ l u t i o n . ~The monomeric state is generally favored by electron-donating groups. The electron-withdrawing nitro group will cause a considerable fraction of the nitronitrosobenzenes to be in the dimeric state V. It is also well known that electronpoor azo groups are excellent dienophiles in DielsAlder reactions.6 The formation of the products I1 may then be accounted for by a reaction between the diene and the electron-poor K=N group of the dimeric nitronitrosobenzene, to form a tetrahydro-l,2-diazine-SJK-dioxide Ar
JANHAMMER AND ROBERT E. BERNARD Department of Chemistry, Tulane University, h'ew Orleans 18, Louisiana
The addition reactions of aromatic nitroso compounds to conjugated dienes are well known2 The reaction products I, dihydrooxazines, generally are isolated in high yields. Side productsof the reaction have not been reported. o-Nitronitrosobenzene' conforms to this pattern, but for the reaction of m- or p-nitrosobenzene with 2 3 dimethyl-1,3-butadiene, we have found two reaction products. One of these reaction products was found to be the expected adduct I. The second reaction product I1 showed an elemental analysis and molecular weight corresponding to an adduct consisting of one mole of diene and two moles of nitroso compound. Aromatic nitroso compounds have been reported to react smoothly with suitably substituted alkenes13 yielding oxazetidines 111. Product 11,therefore, might be the result of the normal 1,4-addition of nitroso com-
Ar-yx-13 Ar
I11
VI. Steric hindrance may explain the failure of o-nitronitrosobenzene to give an adduct of this type, while the low yield of this adduct for m-nitronitrosobenzene may be ascribed to the decreased resonance effect. The infrared absorption band a t about 1050 cni.-' associated with the oxazine ring,2 was absent from the spectra of products 11. Although the spectra of a small number of tertiary amine oxides have been reported6, we have not been able to assign bands in the 970-950-cm. - region unequivocally to the amine oxide. The insolubility of products I1 in suitable solvents ruled out the determination of an nuclear magnetic resonance spectrum. Conclusive evidence for the assertion that products I1 have structure VI was furnished by the deoxygenation of the K,K-dioxide I I a by Ochai's m e t h ~ d em,~ ploying phosphorus trichloride as the deoxygenation agent. The known substance 4,5-dimethyl-1,2-bis(p~-nitrophenyl)-l,2,3,6-tetrahydropyridazines was obtained as the deoxygenation product.
4JAr
0
I
Ar VI
Received October 8, 1962
Ar-YJ0
N-0
v
IV
pound to diene, followed by a 1,2-addition of the nitroso compound to the dihydrooxazine, yielding compound I V or its isomer. I n order to test this hypothesis, we attempted to treat the normal 1,4-adduct of p-nitronitrosobenzene and 2,3-dimethyl-l,3-butadienewith the equivalent amount of p-nitronitrosobenzene. We were unable to detect a reaction with the aid of infrared spectra. Unsuccessful attempts were also made to prepare an oxaxetidine from p-nitronitrosobenzene and the follow(1) This work U A S suppoited b y tlic Petroleum Heseaicll 1 u n d , I t piesented a t the 14th Southeastern Reglonal hleetlng, Gatlinburg, Tenn., November 3,1962. (2) J. Hamer a n d R. E. Bernard. Rec. trav chim , 81, 731 (1962) (3) C. K. Ingold and J. D. Weaver, J. Ckem. Soc., 126, 1146 (1924).
Experimental9 m-Nitronitrosobenzene, m.p. 89-90', and p-nitronitrosobenzene, m.p. 118', were prepared by oxidation with Caro's acid from the corresponding amines.10 Adducts of p-Nitronitrosobenzene.-p-Nitronitrosobensene, 0.50 g. (3.3 mmoles), and 2,3-dimethyl-1,3-butadiene, 0.37 g. (4.5mmoles), were dissolved in 20 ml. of nitromethane a t 0". The green color of the solution, caused by the monomeric nitroso compound, changed in about 15 min. to orange, indicating the (4) B. G. Gowenlock a n d W. Luetkke, Quart. Reu. (London), 12, 321 (1958). ( 5 ) A. Rodgman a n d G. F. Wright, J . Org. Chem., 18, 465 (1953). (6) L. J. Bellamy, "The Infrared Spectra of Complex Molecules," 2nd ed., John U'iley and Sons, Inc., New York, N. Y., 1958, p. 308. (7) E. Ochai, J . Org. Chem.. 18, 550 (1953). ( 8 ) P. Rarangrr, J. Zevisalles, a n d RI. Vuidart, C o m p f . rend., 236, 1365 (1953). (9) All melting points are uncorrected. Analksis by Galbraitli RIirroanalytical Laboratories, Knoxville, Tenn. Infrared spectra measured in potassium bromide with a Beckman IR-5. (10) E. Bamberger and E. Huebner, Ber., 36, 3803 (1803).
